I personally hope that LLMs have no such limits. The good these tools can do is immeasurable. I can already run Llama 2 @70b on my laptop, and that’ll look lik

I personally hope that LLMs have no such limits. The good these tools can do is immeasurable. I can already run Llama 2 @70b on my laptop, and that’ll look like a quaint old AI artifact in 5-7 years. I think the consumer market will keep pace yet stay well below SotA, just as it always has. That still leaves plenty of room for incredible open-source stuff!